Two-Tier Labor Markets in the Great Recession: France vs. Spain - PowerPoint PPT Presentation

Two-Tier Labor Markets in the Great Recession: France vs. Spain Samuel Bentolila Pierre Cahuc CEMFI E. Polytechnique and CREST Juan J. Dolado Thomas Le Barbanchon U. Carlos III E. Polytechnique CPB Netherlands Bureau for Economic Policy Analysis and Research Centre for Education and the Labour Market (ROA) Conference on “Flexibility of the Labour Market” The Hague, 20-21 January 2011 1 / 26

Introduction I Why France and Spain? I Very di¤erent reaction of Spanish unemployment to the crisis relative to France (or EU) Figure 1 I Similar labor market institutions: employment protection legislation, unemployment bene…ts, wage bargaining I BUT: Higher share of temporary labor contracts in Spain: France below 15%, Spain around 33% of employees I What is the role of temporary employment (gap in …ring cost between permanent and temporary jobs) vs. other factors? (residential construction + …nancial crisis) 2 / 26

Introduction I What would have been the evolution of unemployment in Spain if it had French labor market institutions? I In order to answer to this question we I Use a search and matching model with temporary and permanent jobs I Calibrate the model to reproduce the recent evolution of unemployment in France and Spain I Analyze how unemployment would have evolved in Spain if it had French labor market institutions 3 / 26

Introduction I What would have been the evolution of unemployment in Spain if it had French labor market institutions? I In order to answer to this question we I Use a search and matching model with temporary and permanent jobs I Calibrate the model to reproduce the recent evolution of unemployment in France and Spain I Analyze how unemployment would have evolved in Spain if it had French labor market institutions I Result: About 45% of the increase in the unemployment rate would have been avoided had Spain had French institutions; 40% of which is due to its higher …ring costs 3 / 26

Previous literature on temporary jobs in search models I Blanchard and Landier (2002), Cahuc and Postel-Vinay (2002): I Endogenous job destruction w/ temporary and permanent jobs I Temporary jobs ! More job creation and destruction I Bentolila and Saint-Paul (1996), Boeri and Garibaldi (2007): I Transitional honeymoon (job creation followed by reductions in employment) I Sala, Silva, and Toledo (2009): I Calibrated on a representative European labor market; intermediate unemployment volatility I Costain, Jimeno, and Thomas (2010): I Focus on dynamics of unemployment with dual labor market 4 / 26

Our approach I Speci…c event: A negative aggregate shock in France and Spain I Take account of actual features of labor contracts I Temporary jobs cannot be destroyed before their date of termination I Time is needed to destroy permanent jobs I Wages are renegotiated by mutual agreement I Di¤erent types of wage setting: Endogenous or endogenous with …xed bene…ts and …ring cost I Di¤erence-in-di¤erences approach 5 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 6 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 2. Brief comparison of labor market institutions in the two countries 6 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 2. Brief comparison of labor market institutions in the two countries 3. Search and matching model with permanent and temporary contracts 6 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 2. Brief comparison of labor market institutions in the two countries 3. Search and matching model with permanent and temporary contracts 4. Simulations 6 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 2. Brief comparison of labor market institutions in the two countries 3. Search and matching model with permanent and temporary contracts 4. Simulations I To account for the change in unemployment in France and Spain from the boom (2005-2007) to the recession (2008-2009) 6 / 26

Outline 1. Relative performance of the French and Spanish labor markets in the crisis vis-à-vis the expansion 2. Brief comparison of labor market institutions in the two countries 3. Search and matching model with permanent and temporary contracts 4. Simulations I To account for the change in unemployment in France and Spain from the boom (2005-2007) to the recession (2008-2009) I To evaluate what would have been the evolution of unemployment in Spain in the recession if it had French labor market institutions 6 / 26

1. Labor markets before and during the crisis Convergence: A mirage. I Unemployment: France vs. Spain Table 1 I Stronger decrease during the boom I Stronger increase during the recession I Temporary contracts: extreme turbulence I Stock: 14% of employees in France, one-third in Spain (1998) I Job losses over the period 2007:4-2009:4 I France: 276,000 ( 324,000 temporary) I Spain: 1,330,000 (1,380,000 temporary) 7 / 26

2. Labor market institutions 1. Employment protection legislation (EPL) I Firing costs higher in Spain for permanent contracts, lower for temporary contracts, so higher …ring cost gap between permanent and temporary jobs in Spain than in France 2. Unemployment bene…ts I Very similar across countries (taking into account income taxes, entitlement duration rules, and assistance bene…ts) 3. Collective bargaining I Similar in the two countries (Spain copied France in the early 1980s) 8 / 26

2. Labor market institutions 4. Mismatch I A reallocation shock I Construction employment share (2007): France 6.9%, Spain 13.3% Why? Higher fall in real interest rate, ∆ Unskilled labor (HS dropouts, immigrants), Initial size of dual labor market (Saint-Paul, 1997) I Construction was geographically concentrated I Geographical mobility is much lower in Spain I Interregional migration rate: France 2.1%, Spain 0.2% 9 / 26

3. Model setup I Continuum of in…nitely-lived risk-neutral workers and …rms, discount rate r > 0 I Measure of workers = 1 I Matching function à la Pissarides (2000): I m ( u , v ) = m 0 u α v 1 � α I Matching rate for vacancies: q ( v / u ) = q ( θ ) I Matching rate for unemployed: θ q ( v / u ) = θ q ( θ ) I Workers I Unemployed - get unemployment bene…t b ω ( b for short) I Employed on temporary job I Employed on permanent job I Under advance notice 10 / 26

3. Model setup I Job matches with (idiosyncratic) productivity distribution: F ( ε ) � [ ε , ε ] . I ε s Poisson( µ ). I All new jobs start with ε = ¯ ε I When created, a job is I Temporary with probability p I Permanent with probability 1 � p 11 / 26

3. Model setup I Temporary jobs end at rate λ I Either transformed into a permanent job (if their productivity ε if high enough) I or destroyed at zero cost I Permanent jobs I Under advance notice if ε is below an endogenous reservation productivity level I Permanent jobs under advance notice are destroyed at rate σ I Dismissal entails red-tape …ring cost f ω ( f for short) I Wage bargaining on each job: workers get a share β of the surplus I No renegotiation on temporary jobs I Renegotiation on permanent jobs 12 / 26

3. Model setup 13 / 26

3. Comparative statics I Equilibrium Figure 2 I Increase in …ring cost f on permanent jobs: I Firms become I Less strict in …ring permanent workers I More strict in transforming temporary contracts into permanent I Ambiguous e¤ect on I Unemployment I Job destruction (less permanent, more temporary) I Reduction in the probability p of creating temporary jobs: I Less job creation I Less job destruction I Ambiguous e¤ect on unemployment 14 / 26

4. Simulation strategy 1. Solve the model for each country I Parameters: I Calibration (taken from data) I Indirect inference: Match the average rates of unemployment, temporary employment, and permanent job destruction I In two phases of cycle: I Expansion I Recession: Assign values to aggregate productivity shock and (in Spain) to mismatch shock 2. Counterfactual simulation for Spain: With its own shocks (as computed above) and French policy parameters (EPL) 3. Di¤erence-in-di¤erences: Increase in u in Spain with Spanish policy paramters minus Increase in u in Spain with French policy paramters 15 / 26

4. Calibration a) Calibrated parameters: I Environment parameters: I r = 0.01 per quarter I Cobb-Douglas matching function. Hosios: α = β = 0.5 I Institutional parameters ( b , f , p , λ , σ ) Table 2 b) Parameters estimated by indirect inference: I Cost of vacant jobs ( h ) I Matching function scale parameter ( m 0 ) I Job-speci…c productivity shocks arrival rate ( λ ) I Uniformly distributed aggregate shock: γ [ ε , ε ] 16 / 26

4. Simulation results I Matching the data Table 3 I Expansion I Recession: Baseline model (no mismatch shock) and alternative model I Di¤erence-in-di¤erences approach: I u in recession – u in expansion in Spain minus I u in recession – u in expansion in Spain with French policy parameters ( f , p ) I Changes in unemployment I Steady states Table 4 I Transitional dynamics Figure 3 17 / 26